# Predictive Values in Qualitative Assay Validation ## Quick Answer - Predictive values depend directly on disease prevalence in the tested population, so a test with fixed sensitivity and specificity can have very different positive and negative predictive values across clinical settings. - Report predictive values alongside the prevalence estimate used to calculate them, and recalculate them whenever the intended-use population changes. - Predictive values cannot be inferred from sensitivity and specificity alone without knowing prevalence, which is the most common error in validation reporting. ## Understanding Predictive Values in Qualitative Assay Validation Qualitative assays return a binary result, positive or negative, for the presence of a target analyte. Validation of these assays requires laboratory professionals to characterize also how well the test detects the target but also how much confidence a positive or negative result carries in a specific patient population. Predictive values quantify that confidence, and they are the bridge between analytical performance and clinical utility. The central problem in qualitative assay validation is that sensitivity and specificity describe the test itself, while predictive values describe the test in a population. A laboratory can validate an assay with excellent sensitivity and specificity, yet the same assay can produce mostly false positives in a low-prevalence population or mostly false negatives in a high-prevalence population. Understanding this relationship is essential for anyone who interprets validation data, selects assays for clinical use, or reports results to clinicians. This article explains the mathematical relationship between prevalence, sensitivity, specificity, and predictive values. It provides worked examples that laboratory professionals can reproduce with their own data, discusses how to report predictive values in validation studies, and addresses common errors in interpretation. ## At a Glance The table below summarizes the key concepts and their practical implications for qualitative assay validation. | Concept | Definition | Practical Implication | | --- | --- | --- | | Sensitivity | Proportion of true positives correctly identified by the assay | High sensitivity reduces false negatives, which matters when missing a positive result carries high risk | | Specificity | Proportion of true negatives correctly identified by the assay | High specificity reduces false positives, which matters when a false positive leads to unnecessary intervention | | Prevalence | Proportion of the target condition in the tested population | Predictive values cannot be calculated without prevalence, prevalence varies by setting and time | | Positive Predictive Value | Probability that a positive result is a true positive | Decreases as prevalence decreases, even when sensitivity and specificity are fixed | | Negative Predictive Value | Probability that a negative result is a true negative | Increases as prevalence decreases, even when sensitivity and specificity are fixed | | Validation Reporting | Documentation of the population and prevalence used in the study | Predictive values must be reported with the prevalence used to calculate them | ## Core Principles of Predictive Values ### The Relationship Between Prevalence and Predictive Values Predictive values are conditional probabilities. The positive predictive value (PPV) answers the question: given that the assay returned a positive result, what is the probability that the target is actually present? The negative predictive value (NPV) answers the reverse question: given a negative result, what is the probability that the target is actually absent? Both values depend on three inputs: sensitivity, specificity, and prevalence. Sensitivity is the probability that the assay returns a positive result when the target is present. Specificity is the probability that the assay returns a negative result when the target is absent. Prevalence is the proportion of the tested population that actually has the target. The mathematical relationship is expressed through Bayes theorem. The PPV is calculated as the number of true positives divided by the total number of positive results, which includes both true positives and false positives. The NPV is calculated as the number of true negatives divided by the total number of negative results, which includes both true negatives and false negatives. The key insight is that sensitivity and specificity are properties of the assay, while predictive values are properties of the assay in a specific population. A test with 99% sensitivity and 99% specificity will have a PPV of only 50% in a population with 1% prevalence. The same test will have a PPV of 99.9% in a population with 50% prevalence. The assay has not changed, but the meaning of a positive result has changed dramatically. ### Why Prevalence Matters in Validation Prevalence is not a fixed property of the assay. It varies by geographic region, by clinical setting, by patient demographics, and over time. A validation study performed on samples from a high-prevalence referral population will produce predictive values that do not apply to a low-prevalence screening population. Laboratory professionals must therefore define the intended-use population before calculating predictive values. The prevalence used in the calculation should reflect the population in which the assay will be used, not the prevalence in the validation sample set. If the validation sample set is enriched with positive samples, as is common in analytical validation, the prevalence in that sample set will be much higher than the prevalence in the clinical population. Reporting predictive values based on the enriched sample set will overstate the PPV and understate the NPV. The solution is to calculate predictive values using the expected prevalence in the clinical population, not the prevalence in the validation sample set. This requires the laboratory to estimate the prevalence from published literature, local data, or clinical judgment, and to document that estimate in the validation report. ## Worked Examples ### Example 1: High Prevalence Population Consider an assay with 95% sensitivity and 95% specificity, used in a population with a prevalence of 20%. In a cohort of 1,000 patients, 200 have the target and 800 do not. The assay will detect 95% of the 200 true positives, which is 190 true positives. It will miss 5% of the 200, which is 10 false negatives. The assay will correctly identify 95% of the 800 true negatives, which is 760 true negatives. It will incorrectly flag 5% of the 800 as positive, which is 40 false positives. The total number of positive results is 190 true positives plus 40 false positives, which is 230. The PPV is 190 divided by 230, which is 82.6%. The total number of negative results is 760 true negatives plus 10 false negatives, which is 770. The NPV is 760 divided by 770, which is 98.7%. In this population, a positive result is correct about 83% of the time, and a negative result is correct about 99% of the time. ### Example 2: Low Prevalence Population Now consider the same assay with the same sensitivity and specificity, but used in a population with a prevalence of 1%. In a cohort of 1,000 patients, 10 have the target and 990 do not. The sensitivity of 95% detects 95% of the 10 true positives, which is 9.5 true positives. The assay misses 5% of the 10, which is 0.5 false negatives. The specificity of 95% correctly identifies 95% of the 990 true negatives, which is 940.5 true negatives. It incorrectly flags 5% of the 990, which is 49.5 false positives. The total number of positive results is 9.5 true positives plus 49.5 false positives, which is 59. The PPV is 9.5 divided by 59, which is 16.1%. The total number of negative results is 940.5 true negatives plus 0.5 false negatives, which is 941. The NPV is 940.5 divided by 941, which is 99.9%. In this population, a positive result is correct only 16% of the time, even though the assay has the same sensitivity and specificity as in the first example. A negative result is correct more than 99% of the time. ### Example 3: The Effect of Higher Specificity The previous examples show that low prevalence produces many false positives relative to true positives. The solution is not always to increase sensitivity, because sensitivity does not affect the false-positive rate. The solution is to increase specificity. Consider the same 1% prevalence population, but now the assay has 95% sensitivity and 99.5% specificity. In a cohort of 1,000 patients, 10 have the target and 990 do not. The sensitivity of 95% detects 9.5 true positives and misses 0.5. The specificity of 99.5% correctly identifies 985.05 true negatives and incorrectly flags 4.95 false positives. The total positive results are 9.5 plus 4.95, which is 14.45. The PPV is 9.5 divided by 14.45, which is 65.7%. The NPV is 985.05 divided by 985.55, which is 99.9%. Increasing specificity from 95% to 99.5% raises the PPV from 16.1% to 65.7% in the same low-prevalence population. This demonstrates that specificity is the dominant driver of PPV in low-prevalence settings. ## Practical Workflow for Calculating Predictive Values ### Step 1: Define the Intended-Use Population The first step is to define the population in which the assay will be used. This includes the clinical setting, the patient demographics, the geographic region, and the reason for testing. A screening population has a different prevalence than a diagnostic population, and a population with symptoms has a different prevalence than an asymptomatic population. Document the intended-use population in the validation plan. This documentation should include the rationale for the prevalence estimate, whether it comes from published literature, local laboratory data, or clinical judgment. ### Step 2: Estimate the Prevalence The prevalence estimate should be based on the best available evidence. Published studies of the target condition in the relevant population are the preferred source. Local laboratory data can be used if the assay has been in use and the positivity rate has been tracked. Clinical judgment is acceptable when no other data exist, but the basis for the judgment should be documented. The prevalence estimate should be a single number for the calculation, but the report should also include a range of plausible prevalence values. This allows the reader to see how the predictive values change across the range. ### Step 3: Determine Sensitivity and Specificity From Validation Data Sensitivity and specificity are determined from the validation study. The validation study should include a set of known positive and known negative samples, with the true status established by a reference method. The sensitivity is the proportion of known positives that the assay detects, and the specificity is the proportion of known negatives that the assay correctly rejects. The validation study should include enough samples to produce a stable estimate. The number of samples depends on the expected performance of the assay and the desired confidence interval. The laboratory should document the sample size and the confidence intervals for sensitivity and specificity. ### Step 4: Calculate Predictive Values The predictive values are calculated using the prevalence, sensitivity, and specificity. The calculation can be done by hand using the formulas, or with a spreadsheet. The formulas are: PPV = (sensitivity x prevalence) / [(sensitivity x prevalence) + ((1 - specificity) x (1 - prevalence))] NPV = (specificity x (1 - prevalence)) / [(specificity x (1 - prevalence)) + ((1 - sensitivity) x prevalence)] The result should be reported as a percentage with the prevalence used in the calculation. ### Step 5: Report Predictive Values With the Prevalence The report should state the prevalence used in the calculation, the sensitivity and specificity of the assay, and the resulting predictive values. The report should also include the range of predictive values across the plausible range of prevalence. This allows the clinician to interpret the results in the context of their own patient population. ## Options and Tradeoffs in Validation Design ### Enriched Sample Sets Versus Clinical Sample Sets Validation studies often use enriched sample sets that contain a higher proportion of positive samples than the clinical population. This is useful for establishing sensitivity and specificity with a smaller number of samples, because the laboratory does not need to test thousands of negative samples to find a small number of positives. The tradeoff is that the prevalence in the enriched sample set does not reflect the clinical prevalence. If the laboratory calculates predictive values from the enriched sample set, the PPV will be overstated and the NPV will be understated. The laboratory must therefore calculate predictive values using the clinical prevalence, not the sample-set prevalence. ### Reference Method Selection The sensitivity and specificity of the assay are defined relative to a reference method. The reference method should be the best available method for detecting the target, and it should be independent of the assay being validated. If the reference method is imperfect, the sensitivity and specificity of the assay will be biased relative to the true status of the sample. The laboratory should document the reference method and its limitations. If the reference method is known to be imperfect, the laboratory should consider using a composite reference standard or a discrepancy analysis to resolve discordant results. ### Sample Size Considerations The sample size of the validation study determines the precision of the sensitivity and specificity estimates. A small sample size produces wide confidence intervals, which means the predictive values calculated from those estimates are also uncertain. The laboratory should calculate the sample size needed to achieve the desired precision for sensitivity and specificity. The sample size should be reported in the validation report, along with the confidence intervals for sensitivity and specificity. ## Observations and Measurements ### What to Record During Validation The validation study should record the following for each sample: - The sample identifier - The reference method result - The assay result - The sample type and collection date - The operator who performed the assay - The reagent lot and the instrument used These records allow the laboratory to trace any discordant results and to identify sources of variation. The records should be maintained in a format that allows the laboratory to reproduce the validation results. ### How to Measure Prevalence Prevalence is measured by testing a representative sample of the intended-use population with the reference method. The proportion of positive results is the prevalence. This is the most direct measurement, but it is often not feasible because the reference method is expensive or invasive. When direct measurement is not feasible, the laboratory can estimate prevalence from published literature or from local data. The estimate should be documented with the source and the date of the estimate. ### Confidence Intervals The predictive values are point estimates, and they have uncertainty. The uncertainty comes from the uncertainty in sensitivity, specificity, and prevalence. The laboratory should report the confidence intervals for the predictive values, or at least the range of predictive values across the plausible range of prevalence. The confidence intervals can be calculated using standard statistical methods, or the laboratory can use a spreadsheet to calculate the range of predictive values across the range of prevalence. ## Quality Controls and Verification ### Internal Quality Controls Internal quality controls are samples with known results that are tested with each batch of patient samples. They verify that the assay is performing within its established parameters. The controls should include both positive and negative samples, and the results should be recorded. The internal quality controls do not measure predictive values, but they verify that the sensitivity and specificity of the assay have not changed. If the controls fail, the assay results are not valid, and the patient samples should be repeated. ### External Quality Assessment External quality assessment programs provide samples with unknown results to the laboratory. The laboratory tests the samples and reports the results to the program, which compares the results to the expected results. This verifies that the assay is performing correctly across laboratories. External quality assessment does not measure predictive values, but it verifies that the sensitivity and specificity of the assay are consistent with the expected performance. ### Monitoring Prevalence Over Time The prevalence in the clinical population can change over time. The laboratory should monitor the positivity rate of the assay over time to detect changes in prevalence. If the positivity rate changes, the predictive values should be recalculated with the new prevalence. The positivity rate is not the same as the prevalence, because the positivity rate includes false positives. However, a change in the positivity rate can indicate a change in the prevalence, and the laboratory should investigate the cause of the change. ## Common Failure Patterns ### Reporting Predictive Values Without Prevalence A common failure is to report predictive values without stating the prevalence used in the calculation. This makes the predictive values meaningless, because the reader cannot know the population to which they apply. The report should always state the prevalence. ### Using the Validation Sample Prevalence Another common failure is to use the prevalence in the validation sample set to calculate predictive values. This overstates the PPV and understates the NPV, because the validation sample set is often enriched with positives. The laboratory should use the clinical prevalence, not the sample-set prevalence. ### Confusing Sensitivity With PPV Sensitivity is the probability that a positive sample is detected, while PPV is the probability that a positive result is a true positive. These are different concepts, and they are often confused. A test with high sensitivity can have a low PPV in a low-prevalence population. ### Ignoring the Effect of Specificity In low-prevalence populations, the PPV is driven by specificity, not sensitivity. A test with high sensitivity but low specificity will produce many false positives in a low-prevalence population. The laboratory should focus on specificity when the prevalence is low. ## Limitations and Interpretation ### The Reference Method Problem The sensitivity and specificity of the assay are defined relative to the reference method. If the reference method is imperfect, the sensitivity and specificity of the assay are biased. The predictive values calculated from these biased estimates are also biased. The laboratory should document the reference method and its limitations. If the reference method is imperfect, the laboratory should consider using a reference standard or a discrepancy analysis. ### The Prevalence Estimation Problem The prevalence estimate is often based on published literature or clinical judgment, and it may not reflect the actual prevalence in the local population. The laboratory should document the source of the prevalence estimate and the range of plausible values. ### The Generalizability Problem The predictive values are specific to the population and the prevalence used in the calculation. They do not generalize to other populations with different prevalence. The laboratory should report the predictive values for the intended-use population, and the clinician should interpret the results based on the prevalence in the patient population. ## Biosafety and Regulatory Context ### Biosafety in Validation Studies Validation studies involve the handling of biological samples, which may contain infectious agents. The laboratory should follow the biosafety principles in the Biosafety in Microbiological and Biomedical Laboratories, including the risk assessment, the containment, and the institutional oversight. The laboratory should use the appropriate biosafety level for the samples being handled. The laboratory should also follow the NIH Guidelines for Recombinant or Synthetic Nucleic Acid Molecules if the validation involves recombinant or synthetic nucleic acids. These guidelines provide the oversight for the use of recombinant and synthetic nucleic acid molecules. ### Quality Standards The laboratory should follow the quality standards for the measurement science, which are established by the National Institute of Standards and Technology. These standards provide the basis for the calibration and the quality of the measurements. The laboratory should also follow the MIQE guidelines for the reporting of qPCR experiments. The MIQE guidelines provide the minimum information for the reporting of qPCR experiments, including the controls, the transparency, and the interpretation. ## Professional Escalation Criteria The laboratory should escalate the following situations to the appropriate professional: - The predictive values are below the acceptable threshold for the intended use - The prevalence in the clinical population changes significantly from the prevalence used in the validation - The reference method is found to be imperfect, and the sensitivity and specificity are biased - The internal quality controls fail repeatedly - The external quality assessment results are unacceptable The escalation should include the documentation of the issue, the data, and the proposed action. ## Decision Framework for Prevalence-Adjusted Reporting in Routine Diagnostic Work Laboratory professionals who complete a validation study face a second problem that the initial calculations do not solve: how to keep predictive values useful after the assay enters routine service. The prevalence used in the validation report is a snapshot taken at one point in time, and the population tested by the laboratory can shift with seasonal disease patterns, referral changes, new screening programs, or changes in the local patient base. A decision framework that ties predictive value reporting to ongoing prevalence monitoring gives the laboratory a defensible method for updating reports without redoing the full validation study. ### The Prevalence Monitoring Decision Cycle The framework operates on a defined cycle with four stages: baseline establishment, scheduled monitoring, threshold-triggered recalculation, and report revision. Each stage has specific actions, records, and escalation criteria. **Stage 1: Baseline Establishment** The baseline is the prevalence estimate used in the initial validation report. This estimate comes from the sources described in the validation workflow, which include published literature, local laboratory data, or clinical judgment. The baseline must be recorded with its source, the date of the estimate, and the rationale for choosing that value over other available estimates. The baseline also includes the range of plausible prevalence values that the laboratory identified during validation. This range becomes the reference band for future monitoring. For example, if the validation report used a prevalence of 8% with a plausible range of 5% to 12%, the reference band is 5% to 12%. The laboratory does not need to recalculate predictive values as long as the monitored prevalence remains inside this band. **Stage 2: Scheduled Monitoring** The laboratory monitors the assay positivity rate on a fixed schedule. The schedule should be defined in the validation plan and should reflect the testing volume of the laboratory. A high-volume laboratory that runs hundreds of tests per week can monitor monthly. A low-volume laboratory that runs dozens of tests per month should monitor quarterly or after every 100 consecutive results, whichever comes first. The monitoring record for each period includes the number of tests performed, the number of positive results, the positivity rate, and the date of the monitoring period. The positivity rate is not the same as the prevalence, because the positivity rate includes false positives. However, the positivity rate is the only rate that the laboratory can observe directly during routine service. The laboratory uses the positivity rate as a proxy for prevalence and applies the relationship between the two to interpret changes. The relationship between the positivity rate and the prevalence is defined by the assay sensitivity and specificity. The expected positivity rate is calculated as the sum of the true positives and the false positives in a hypothetical cohort. For an assay with 95% sensitivity and 95% specificity, the expected positivity rate at 1% prevalence is the true positive rate of 0.95% plus the false positive rate of 4.75%, which is 5.7%. At 20% prevalence, the expected positivity rate is 19% plus 4%, which is 23%. The laboratory can calculate the expected positivity rate for the baseline prevalence and compare the monitored rate to that expectation. **Stage 3: Threshold-Triggered Recalculation** The laboratory recalculates the predictive values when the monitored positivity rate moves outside the reference band. The reference band is not the same as the plausible range of prevalence. The reference band is the range of positivity rates that corresponds to the plausible range of prevalence, adjusted for the assay sensitivity and specificity. The laboratory calculates the upper and lower bounds of the reference band by applying the sensitivity and specificity to the upper and lower prevalence bounds. For an assay with 95% sensitivity and 95% specificity, and a prevalence range of 5% to 11%, the lower bound of the positivity rate is the expected rate at 5% prevalence, which is 9.5% true positives plus 4.75% false positives, or 14.25%. The upper bound is the expected rate at 11% prevalence, which is 10.45% plus 4.45%, or 14.9%. The reference band is 14.25% to 14.9%. If the monitored positivity rate falls outside this band, the laboratory recalculates the predictive values using the new prevalence estimate. The new prevalence estimate is derived from the monitored positivity rate, adjusted for the assay sensitivity and specificity. The laboratory can solve for the prevalence using the observed positivity rate and the known sensitivity and specificity. The formula for the prevalence is the observed positivity rate minus the false positive rate, divided by the difference between the sensitivity and the false positive rate. The laboratory should not use a single monitoring period to trigger a recalculation. A single period can be affected by random variation, especially in low-volume laboratories. The framework requires two consecutive monitoring periods outside the reference band before the laboratory recalculates the predictive values. This requirement reduces the risk of overreacting to random fluctuation. **Stage 4: Reporting Revision** When the laboratory recalculates the predictive values, it must revise the validation report. The revised report includes the new prevalence estimate, the date of the recalculation, the monitoring data that triggered the recalculation, and the new predictive values. The revised report also includes the previous prevalence and predictive values, so the reader can see the change over time. The revised report should be distributed to the same audience as the original validation report. This includes the laboratory director, the clinical staff who order the assay, and any institutional review or quality committees. The laboratory should document the distribution of the revised report. ### Record System for Prevalence Monitoring The record system for prevalence monitoring is separate from the validation study records. The monitoring records are maintained in a format that allows the laboratory to track the positivity rate over time and to demonstrate the basis for any recalculation. The monitoring record for each period includes the following fields: - Monitoring period start and end dates - Number of tests performed in the period - Number of positive results in the period - Positivity rate for the period - Expected positivity rate at the baseline prevalence - Whether the period is inside or outside the reference band - The operator who compiled the data - The date the data were reviewed The laboratory should also record the cumulative data. The cumulative positivity rate over a longer period is more stable than the rate for a single period, and it can be used to confirm a trend that appears in the period data. The monitoring records should be reviewed by the laboratory director or the designated quality professional on the same schedule as the monitoring. The review should be documented with the reviewer name and the date. ## Troubleshooting Method for Prevalence Changes When the monitored positivity rate moves outside the reference band, the laboratory must determine whether the change is due to a true change in prevalence, a change in assay performance, or a change in the testing population. The troubleshooting method uses a sequence of checks that isolate the cause. **Check 1: Assay Performance** The first check is the internal quality control data. If the internal quality controls show a change in the control values, the assay performance may have changed. The laboratory should review the control records for the monitoring period and compare them to the established control ranges. If the controls are out of range, the assay is not performing within its validated parameters, and the predictive values are not valid. The laboratory should investigate the cause of the control failure before interpreting the monitoring data. **Check 2: Sample Mix** The second check is the composition of the tested population. The laboratory should review the test requisitions for the monitoring period to determine whether the patient mix has changed. A new referral pattern, a new screening program, or a change in the ordering clinician can change the prevalence in the tested population without any change in the assay. The laboratory should document any changes in the population and consider whether the new population matches the intended-use population from the validation. **Check 3: Reference Method Stability** The third check is the reference method. If the reference method has changed, the sensitivity and specificity of the assay are no longer defined relative to the same standard. The laboratory should verify that the reference method has not changed since the validation. If the reference method has changed, the laboratory should repeat the sensitivity and specificity determination before recalculating the predictive values. **Check 4: Prevalence Recalculation** If the first three checks show no change in assay performance, population mix, or reference method, the laboratory concludes that the prevalence has changed. The laboratory recalculates the prevalence from the monitored positivity rate and updates the predictive values. The troubleshooting method is documented in the monitoring record. The laboratory records the results of each check, the conclusion, and the action taken. ## Comparison of Reporting Approaches The decision framework offers a structured alternative to two common reporting approaches. The first approach is to report predictive values once at the time of validation and never update them. This approach is simple but becomes inaccurate as the prevalence changes. The second approach is to recalculate predictive values continuously from the monitored positivity rate. This approach is responsive but can produce unstable estimates in low-volume laboratories because the positivity rate fluctuates from period to period. The decision framework sits between these two approaches. It uses the reference band to absorb the normal fluctuation in the positivity rate and only triggers a recalculation when the change is large enough to be meaningful. This reduces the number of recalculations while still keeping the predictive values current. The framework also provides a documented basis for the recalculation. The laboratory can show the monitoring data, the reference band, and the trigger that led to the new predictive values. This documentation is useful for quality audits and for clinical users who need to understand why the predictive values changed. ## Records and Measurements for the Framework The laboratory should maintain the following records for the prevalence monitoring framework: - The validation report with the baseline prevalence and the plausible range - The monitoring table with the period data - The reference band calculation - The troubleshooting records for any period outside the reference band - The revised validation reports with the new predictive values The records should be retained for the same period as the other validation records, which is typically the life of the assay plus the institutional retention period. The records should be available for review by the laboratory director, the quality committee, and any external auditors. The laboratory should also track the number of recalculations per year. A high number of recalculations indicates that the prevalence in the tested population is unstable, and the laboratory should consider whether the intended-use population is correctly defined. A low number of recalculations indicates that the prevalence is stable and the baseline estimate is accurate. ## Professional Escalation Criteria for the Framework The laboratory should escalate the following situations to the laboratory director or the appropriate professional: - The monitored positivity rate moves outside the reference band for three consecutive periods, indicating a sustained change in the population - The troubleshooting check identifies a change in the reference method that requires a new sensitivity and specificity determination - The internal quality controls fail repeatedly, indicating a problem with the assay that cannot be resolved by routine maintenance - The recalculation produces predictive values that fall below the acceptable threshold for the intended clinical use The escalation should include the monitoring data, the troubleshooting records, and the proposed action. The laboratory director decides whether to update the validation report, to change the intended-use population, or to stop the assay pending further investigation. The framework gives the laboratory a practical method for keeping predictive values current without redoing the validation study. It uses the monitoring data that the laboratory already collects, and it provides a documented basis for the updated predictive values. The framework is a distinct addition to the validation workflow, and it addresses the problem of prevalence changes that the initial validation cannot predict. ## Frequently Asked Questions ### What is the difference between sensitivity and positive predictive value? Sensitivity is the probability that the assay detects the target when the target is present. Positive predictive value is the probability that a positive result is a true positive. Sensitivity is a property of the assay, while positive predictive value is a property of the assay in a specific population. ### Why does prevalence affect predictive values? Prevalence affects the number of true positives and true negatives in the population. In a low-prevalence population, the number of false positives can exceed the number of true positives, which lowers the positive predictive value. In a high-prevalence population, the number of false negatives can exceed the number of true negatives, which lowers the negative predictive value. ### How do I calculate the positive predictive value? The positive predictive value is calculated as the number of true positives divided by the total number of positive results. The total number of positive results is the sum of the true positives and the false positives. The calculation requires the sensitivity, the specificity, and the prevalence. ### What is the difference between prevalence and the positivity rate? Prevalence is the proportion of the population that has the target, as determined by the reference method. The positivity rate is the proportion of the population that tests positive with the assay. The positivity rate includes false positives, so it is higher than the prevalence when the specificity is less than 100%. ### How should I report predictive values in a validation report? The report should state the prevalence used in the calculation, the sensitivity and specificity of the assay, and the resulting predictive values. The report should also include the range of predictive values across the plausible range of prevalence. ### What is the effect of specificity on the positive predictive value? In a low-prevalence population, the positive predictive value is driven by the specificity. A small decrease in specificity can produce a large increase in the number of false positives, which lowers the positive predictive value. The laboratory should focus on the specificity when the prevalence is low. ### Can I use the prevalence in the validation sample set to calculate predictive values? No. The prevalence in the validation sample set is often enriched with positive samples, which overstates the positive predictive value and understates the negative predictive value. The predictive values should be calculated using the prevalence in the clinical population. ### What should I do if the prevalence changes after the validation? The predictive values should be recalculated with the new prevalence. The laboratory should monitor the positivity rate over time to detect changes in prevalence, and the predictive values should be updated when the prevalence changes. ## Using the Evidence | Source | Best use in this topic | Important limitation | |---|---|---| | [Molecular Biology and Laboratory Methods](https://www.ncbi.nlm.nih.gov/books) | official guidance | Check the linked page for current local requirements | | [Biosafety in Microbiological and Biomedical Laboratories](https://www.cdc.gov/labs/bmbl/index.html) | official guidance | Check the linked page for current local requirements | | [NIH Guidelines for Recombinant or Synthetic Nucleic Acid Molecules](https://osp.od.nih.gov/policies/biosafety-and-biosecurity-policy/nih-guidelines-for-research-involving-recombinant-or-synthetic-nucleic-acid-molecules) | official guidance | Check the linked page for current local requirements | ## Related Diagnostic Guides - [Method Validation in Clinical Laboratories: A Practical Protocol](/knowledge/diagnostics/emerging-tech/method-validation-in-clinical-laboratories-a-practical-protocol) - [Gram-Positive vs Gram-Negative Cell Walls: A Comparative Guide](/knowledge/diagnostics/emerging-tech/gram-positive-vs-gram-negative-cell-walls-a-comparative-guide) - [Positive and Negative Controls in Diagnostic Assays: Selection and Interpretation](/knowledge/diagnostics/emerging-tech/positive-and-negative-controls-in-diagnostic-assays-selection-and-interpretation) - [Gram-Positive vs. Gram-Negative Cell Walls: Structural Differences and Diagnostic Implications](/knowledge/diagnostics/emerging-tech/gram-positive-vs-gram-negative-cell-walls-structural-differences-and-diagnostic-implications) ## References and Further Reading - [Molecular Biology and Laboratory Methods](https://www.ncbi.nlm.nih.gov/books). National Library of Medicine. - [Biosafety in Microbiological and Biomedical Laboratories](https://www.cdc.gov/labs/bmbl/index.html). CDC and NIH. - [NIH Guidelines for Recombinant or Synthetic Nucleic Acid Molecules](https://osp.od.nih.gov/policies/biosafety-and-biosecurity-policy/nih-guidelines-for-research-involving-recombinant-or-synthetic-nucleic-acid-molecules). National Institutes of Health. - [Addgene Protocols](https://www.addgene.org/protocols). Addgene. - [MIQE Guidelines](https://academic.oup.com/clinchem/article/55/4/611/5631762). Clinical Chemistry. - [NIST Laboratory Programs](https://www.nist.gov/laboratories). National Institute of Standards and Technology. - [Laboratory Biosafety Manual](https://www.who.int/publications/i/item/9789240011311). World Health Organization. - [NCBI Literature Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information. - [Does Past Myopia Progression Predict Future Progression?](https://pubmed.ncbi.nlm.nih.gov/41543335). Investigative ophthalmology & visual science, 2026. - [Determination of a Positive Predictive Value for Confirming HBsAg Qualitative II on Architect and Alinity i (Abbott Diagnostics).](https://pubmed.ncbi.nlm.nih.gov/40779480). Clinical laboratory, 2025. - [PREDICTIVE VALUE OF THE GENERAL MOVEMENTS ASSESSMENT IN PRETERM INFANTS: A META-ANALYSIS.](https://pubmed.ncbi.nlm.nih.gov/32401947). Revista paulista de pediatria : orgao oficial da Sociedade de Pediatria de Sao Paulo, 2020. > This article is educational and does not replace validated laboratory procedures, institutional biosafety review, manufacturer instructions, or professional interpretation.